Gudhem the toponymic evidence (or rather challenge)

Transcription

1 Gudhem the toponymic evidence (or rather challenge) By Stefan Brink, Aberdeen Keywords: place names, pagan cult sites, Scandinavia Abstract: In this article the debate about Gudhem/Gudme place names in Scandinavia is reviewed and evaluated, especially the analyses made by the onomasticians Jöran Sahlgren and John Kousgård Sørensen. In an attempt to understand these sites and their names, potential cult or sacred sites of different kinds are listed and discussed. The main conclusion reached is that we are not dealing with proper names, Gudhem etc., but rather with an underlying appellative, gudhem (goðheimr), the original meaning of which is very difficult to grasp. Why and how this appellative was used to form the place names cannot be determined. In Scandinavia, there are eleven Gudhem (Gudme etc.) place names. An analysis of these names and the places so named has revealed several linguistic, religious and spatial problems. Any attempt to put them all under one umbrella is, as we shall see, a huge challenge. The name Gudhem was highlighted and discussed by two prominent Scandinavian onomasticians, Jöran Sahlgren (1954) and John Kousgård Sørensen (1985), and the discussion was later summed up by Heinrich Beck (1995). After collecting and analysing the names, Sahlgren came to the rather ingenious conclusion that they all had a Christian background, whereby two were original and the rest were derived names (Sw uppkallelsenamn ) from one or other of these two. He argued that the inspiration and precedent for this came from certain famous European monastery and nunnery names, such as Lieu-Dieu, Le Mont Dieu, Gottesthal, Gotteszell etc. The Gudhem monastery and nunnery names in Scandinavia were, according to Sahlgren, translations of Locus Dei or Domus Dei. The problem with Sahlgren s interpretation, as he obviously realised himself a few years later (Sahlgren 1958) or at least led him to have second thoughts and modify his opinion is of a linguistic order. All the continental names of this kind are genitival compounds, as are for example the Scandinavian monastery name Gudsberga (Mons Domini) in Dalarna, Sweden. Hence, another explanation had to be found and a new interpretation was suggested in 1985 by John Kousgård Sørensen, in what has been hailed as a brilliant study. That may be true, but it left several loose ends and problems that were not solved in the article. But let us first reiterate Sahlgren s arguments, which, as I stated, were rather ingenious and must be kept in mind when dealing with these names. He starts by pointing out that Goðheimr is a name mentioned several times by Snorri in his Ynglinga saga, as is the name Manheimar, the homes of humans. (Of course, we do not know to what extent Snorri was building on earlier traditions or just making up 15

2 a good story himself.) In chapter 9, especially, Óðinn is said to have gone to Goðheimr when he died, to meet his friends, hence the dead warriors 1. Sahlgren assumes that Snorri got Manheimar from Eyvindr skáldaspillir and his poem Hákonarmál, but he thinks the idea of Goðheimr comes from Egill s Sonatorrek: Þat man ek enn er upp um hóf í goðheim Gauta spjalli ættar ask þann er óx af mér ok kynvið kvánar minnar. (Egils saga, ed. Bjarni Einarsson 2003) I still do remember, that Gauti s friend [i.e. warrior of Óðinn] was uplifted to Goðheimr. Scion of my family my own seed and my wife s stem (my own free translation) Sahlgren states that the tradition of Óðinn going to Goðheimr obviously existed already at the time of Egill (10 th century) and he concludes that the Icelandic Goðheimr (Goðheimar) is only known as an abode for gods and the deceased, never as a place in the real world. Sahlgren then analyses the names and the corresponding places. Gudhem in Västergötland is the name of a nunnery mentioned in a letter dated 1175 as in Gudhem (SD 1 p. 97). The traditional interpretation of this name is that it contains the words gud pagan god and hem farm, habitation. However, Sahlgren finds it odd that two nunneries of the Benedictine order, in Sweden and in northwestern Jutland in Denmark, were established more or less simultaneously at places that were both already called Gudhem. This leads him to assume that both are translations of Domus Dei or Locus Dei, and one has to admit that Sahlgren has good reasons for this assumption. He then takes the idea a step further; that the Gudhem nunnery in Västergötland got its name in the same way as German monasteries with names beginning with Gott-, and in connection (as a precedent) with the neighbouring farm of Tunhem, which he thinks was the original farm in the area, and which gave land for the nunnery. Regarding the three Norwegian Gudum, Gudim (< Guðheimr), he concludes that the medieval Gudum at Skiptvet was owned by Skiptvet church, Gudim at Rakkestad was owned by Rakkestad church and Gudum at Botne by Botne church. According to Sahlgren, the fact that these three farms, the only Gudhem names found in Norway, all belonged to churches cannot be coincidental ( Att blott kyrkohemman i Norge fått namnet Gudhem, kan knappast bero på en tillfällighet, Sahlgren 1954, 4). This leads him to suppose they were named after one or other of the two nunneries. Gudhem at Veta in Östergötland was a small farm, owned by a noble ( ett frälsehemman om 1 oförmedlat mantal ). Extrapolating from several farms owned by nobles that were given derived names, such as Rosendal, he argues that Gudhem at Veta should also be understood in the same way. He believes that Gudhjem on Bornholm must have been named after Gudum nunnery on Jutland, an opinion he continued to hold in The fact that Gudme on Fyn is the only hem name on Fyn is the argument for this name being a derived name, too. Finally, Sahlgren notes that Gudum at Slagelse on 1 For references to and quotes from Snorri s Ynglinga saga in relation to the name and place Goðheimr, see Lydia Klos article concerning Literature in this vol. 2 Att namnet på fiskeläget Gudhjem på Bornholm skulle vittna om hednisk gudstro, finner jag alldeles osannolikt. Enligt min tanke är dess namn lånat från ett av de nordiska klostren., i.e. That the name of the fishing village Gudhjem on Bornholm could indicate pagan faith, I find totally impossible. I am of the opinion that its name was borrowed from one of the Nordic monasteries (Sahlgren 1958, 187). 16

3 Sjælland was owned by Sorø monastery, which became insolvent in the 12 th century. When, in 1162, Archbishop Absalon handed the monastery over to the Cistercians, together with numerous farms, this might have been the occasion, in Sahlgren s opinion, for the Cistercians to rename the monastery as the former name perhaps had inappropriate associations, in the same way that Ards (< Ars), belonging to Vadstena monastery, was renamed Rosendal for obvious pious reasons. Sahlgren sums up his analysis thus: of the nine Scandinavian Gudhem, three were old monasteries/ nunneries or monastery manors owned by the Cistercians or the Benedictines, three were farms owned by the Church, one was owned by a noble. The remaining two are in areas with no other hemnames. Sahlgren therefore sees strong reasons to believe that the Scandinavian place name Gudhem is related to the Christian God. As mentioned above, one has to admit that Sahlgren had good arguments for such an assumption, but one thing he forgot to do was to analyse the name and the word Goðheimr from a linguistic point of view, a remarkable omission for this rock-solid philologist: why the stem compound here and not, as in for example Gudsberga, a genitive compound. Obviously Sahlgren became aware of this crucial point later and therefore, while not completely discarding his earlier interpretation, modified his position. He is still of the opinion that the Gudhem names are most probably derived names ( uppkallelsenamn ), and agrees that it seems improbable that two nunneries of the Benedictine order, founded simultaneously, were built at places which both had the older name Gudhem. This is not likely, he states. The modification of his opinion is contained in a rather obscure sentence that is difficult to interpret: Men givetvis kan ett inhemskt Gudhem ha inverkat på översättningen av Locus dei, i.e. But, of course, a Gudhem homestead may have influenced the translation of Locus Dei (Sahlgren 1958, 187). Now, let us turn to the analysis by John Kousgård Sørensen (1985). After presenting the material, i.e. eleven names with their medieval written evidence, Kousgård Sørensen discusses and dismisses Sahlgren s interpretation before giving his own. The first element, he concludes, is almost certainly to be understood as the word ON guð n. (pagan) god or pagan gods. This originally neutral word has semantic cognates in words like ON regin, bond, hopt, all of which also take the neutral gender. The names of individual gods found in place names are genitival compounds, such as Torsberg, Odensvi, Fröslunda etc., which makes it more than probable that the Gud- in Gudhem is to be understood as the plural guð pagan gods. But why is this word not found as a normal genitival compound, *Guða-? This is perhaps what would be expected although it is not discussed by Kousgård Sørensen; nevertheless he proposes a solution. In Egill Skallagrímsson s famous lamentation over his dead son, the aforementioned poem Sonatorrek, we find í goðheim, which certainly is to be translated as the home of the gods. As we have seen, the word is also found in Snorri s Ynglinga saga as Goðheimar, the homes of the gods. This is not a genitival compound but rather an ancient stem compound of the same kind as is found in other real or mythical names such as Manheimar the homes of humans, Ásgarðr the home of the ása gods, Danmark the mark of the Danes, Gotland the land of the Gutar, Gautland the land of the Götar, Svíþjóð the (land of the) Svear, Jylland (< Iutland) the land of the Jutar etc. A characteristic of these names is that the first element is the designation of a people (hence the plural) or euhemeristically transformed gods. The next problem is to understand this compound, Gudhem. Was it originally a word or a name? Kousgård Sørensen shows that five of the Gudhem names are found outside the distribution area of the Scandinavian -hem/-heim names, which is a very strong argument for assuming that these names are not real place names ending in -hem/-heim, but have their origin in an appellative, which would thus have to be the appellative found in the Old Norse sources, as in the word guðheimr. The -hem/-heim ending found in these place names (and the appellative) is hence not the place-name element -hem/-heim, but the word hem/heim. This word has a close cognate in the Greek keîmai lies, and an analysis of the word and the place-name element in Scandinavian reveals that there are two possible meanings: 17

4 home, as in Sæheim the farm by the lake/sea, and settlement district, as in Trondheim; meanings that are gradually merging and creating a semantically blurred picture (cf. Brink 1991; 1995). It is important to remember this, although it complicates matters somewhat, when discussing the denotation of the Gudhem names. However, since the more spatially extended meaning as a settlement district is mainly found in very old place names, it seems preferable in the gudhem appellative to consider either the meaning home, farm or the word hem, used here with a derived meaning world. Kousgård Sørensen concludes that the appellative gudhem should be translated as the place where the gods sojourn, their home and where they are the object of a particular cult, which, in my opinion, is debatable. The meaning of gudhem is best explained by Oluf Rygh in his interpretation of the name Gudvin: vel betegne Stedet, som paa en eller anden Maade staaende i Forbindelse med den hedenske Gudstjenste (i.e. probably denoting a place, which is in some way connected with pagan sacred rites ). However, if this assumption is valid, one would certainly expect the names of these locations to denote the actual site of the cult, or of the ritual activity performed, rather than the more elusive Gudhem home of the gods. A better and more precise definition of the appellative gudhem is certainly needed. In our search for answers, we are therefore forced to extend our analysis of the Gudhem problem, away from being merely a philological problem, to involve evidence from other disciplines. A first step is to put the Gudhem settlements into a situational and landscape context, another since the word/name Gudhem has an obvious cultic-religious dimension is to try to decide what kind of cultic context surrounds these sites. An important and interesting point is the distribution of the Gudhem names (cf. Fig. 1 on p. 26). They occur in Denmark, on Bornholm, in south-western Sweden and in Viken, in Øst- and Vestfold. This is more or less exactly the same distribution as, for example, the Scandinavian -lev-/-löv-names, and also some other place names typically found in Denmark and southern Scandinavia (cf. Brink 1997, 76). This is also the same area that, in the Viking Age, is dominated by archaeological artefacts of often Danish origin, an area identified by archaeologists as belonging to the Danish cultural sphere during that period. It is therefore clear that the Gudhem names, although very few, have a distinct distribution in Scandinavia, in a Danish or southern Scandinavian cultural context, i.e. they are only found in Denmark and in areas formerly belonging to or dominated by the Danes, but never in the other two important cultural regions in Scandinavia of the time, the old Svea territory and the old Norse territory, in western, central and northern Norway or Iceland. The only exception seems to be Gudhem in Östergötland. This can, however, be explained by the fact that the eastern part of this province displays a distinct and well-attested Svea influence, whereas the westernmost part, including Lake Vättern where Gudhem is located, was clearly influenced by and affiliated with the province of Västergötland. In other words, the distribution area of Gudhem names can probably be seen as belonging to a Danish or southern Scandinavian cultural sphere. There seems to be no sign of these names in eastern, northern and western Scandinavia. To take a simple and unbiased starting point, the following types of ritual or cultic site seem to be relevant and may be applicable here: 1) places of mythical importance for a society, where mythical persons accomplished some special deed; 2) places where a god or gods performed some act or revealed its/their existence; 3) locations of buildings used for ritual activities, such as offerings; 4) places looked upon as entrances to an Other World, which facilitate contact and communication with the gods (or ancestors), e.g. via offerings; 5) places where spiritual guidance or wisdom can be obtained from the gods (or ancestors), or which allow a person to become spiritually charged ; 18

5 6) sacredly charged places used for rituals, initiation rites, feasts, dancing, etc.; 7) areas or districts dedicated to, or assumed to be controlled by, a god or gods; 8) places of worship and cult, perhaps with some representation of the deity/deities, such as an idol, staff(s), mountain, etc.; 9) other undefined spaces where deities are assumed to dwell or appear. Such sites exist in most ethnic cultures. For example, sites where some mythical (or semi-mythical) individual god or human performed a specific act or deed, was revealed, or was killed (nos. 1, 2 and 9) are linked with a large number of Maori legends (i.e. their history!) and commemorated in Maori place names (Wilson et al. 1990a; 1990b). Special buildings (no. 3), somewhat similar to Greek and Roman temples (but definitely not temples!), are also found in Scandinavia at places such as Uppåkra, Sanda, Borg etc., and have been interpreted as pagan cult buildings (see Jørgensen 2009, passim). In many cultures, sources of water wells, lakes, bogs, rapids and rivers have been looked upon as entrances or interfaces between the human world and the supernatural world, and the same has been assumed for grave mounds, burial grounds, hilltops and groves (no. 4). Certain sites in the landscape where one can acquire wisdom or become spiritually charged are often the location of Thing assemblies, kava assemblies (Polynesians) or dancing grounds (Aborigines) (see i.a. Amituana i 1986, 37 and passim) (no. 5 and 6). Open air cult sites, with or without representations of the worshipped deity, are the probable origin of the vi/vé and harg/horgr sites in Scandinavia, as well as of sites with names containing the elements lund, stav etc. (see Vikstrand 2001, passim) (no. 3, 5 and 6). A particularly interesting case exists when it is possible that a whole district or landscape was identified with a deity or had some significant sacral dimension. Such sacred landscapes have been discussed occasionally in recent years, and some early Scandinavian administrative districts, such as Frøs herred in Denmark and Frösthult (< Frøstolpt) in Sweden (Brink 2007, 111) (no. 7) have long been observed and discussed in this connection. Finally, there are the more elusive areas, with rather undefined borders, where a deity or a god is assumed to appear, dwell, or exist, such as a forest, a grove, a lake, an island, etc. (see Brink 2001, passim) (no. 9). When such sites have been thoroughly analysed in conjunction with archaeological excavations, the observed time depth has very often astonished the excavators, making them hesitate in their interpretations. To assume cult continuity over many hundreds of years, sometimes over a thousand years, undoubtedly makes archaeologists nervous and hesitant (e.g. Berta Stjernquist, Kalle Sognnes; see Brink 2001, 106). However, the accumulation of evidence forces us to acknowledge an astonishing time depth in many cases, sometimes in spite of presumed religious shifts. I have discussed the background to this phenomenon elsewhere (Brink 2001), and would like to suggest that mythical sites and cult places should be looked upon as a kind of investment in the landscape, which create very stable entities that are not easily changed, unless we can assume a total depopulation of the region. Another important aspect to be born in mind is the pagan approach to the landscape in early, pre- Christian Scandinavia, as has been extensively and ingeniously discussed for Chinese and Greek cultures by the American Geographer Yi Fu-Tuan (e.g. 1974, 1977). In an unpublished study, I observe that for pagan Scandinavians the whole landscape was charged in different ways: objects and spaces in the landscape were identified with specific genders (rivers were feminine, lakes masculine, arable land feminine, islands feminine, mountains masculine, etc.). Similarly, the landscape also seems to have been sacredly charged in different ways, with some objects or spaces having greater sacral significance than others. For example, a mountain could be assumed to be the abode of a god, or an ancestor, as in the famous episode of Þórólfr Mostrarskegg. In both the Landnámabók and the Eyrbyggja saga we learn of the landnám by this Þórólfr Mostrarskegg. When he had chosen land to settle on in Iceland, he had to helga sér land, i.e. to perform a ritual to claim the land and, with some kind of magic, make the land heilagr ( holy ) for himself and his heirs. 19

6 Þórólfr s settlement was on a large peninsula called Þórsnes the peninsula dedicated to the god Þórr. On the peninsula was a mountain, a fjall: á þvi fjalli hafði Þórólfr svá mikinn átrunad at þangat skyldi engin maðr óþveginn líta þat fjall kallaði hann Helgafell ok trúði at hann mundi þangat fara, þá er hann dœi, ok allir á nesinu hans frændr 3. On this peninsula, Þórólfr built a temple dedicated to the god Þórr and also created a thing assembly site. From this episode in the saga, we may assume that the authors of the sagas knew or assumed that certain lands and particular physical features in the landscape were charged with metaphysical energy or godly power, or that god(s) supposedly dwelt there; in this case the mountain was therefore given the epithet heilagr holy (Helgafell). But what is a home of the gods especially if it is possibly the name of a settlement or farm? One way to find an answer to this puzzle may be to analyse other cult sites, places of cultic-religious importance and areas or districts identified with a god or gods in the pre-christian landscape of Scandinavia, to see if there are any which might be of interest for our purpose. This field of research is huge and impossible to cover in this article. For anyone interested, the best overview, with in-depth analyses, is to be found in Per Vikstrand s thesis Gudarnas platser (2001). As mentioned above, small buildings have been excavated in Scandinavia, which the excavators have interpreted as cult buildings; historians of religion have accepted this interpretation. It is probably possible to link these sites to place-name elements such as harg/horgr and perhaps also hov/hof. Quite recently we have also excavated obvious cult sites in Sweden with names containing the elements lund(a) and vi, which are to be interpreted as some kind of open-air cult sites (see Vikstrand 2001, passim). Two of the vi sites, in particular, have elaborate platform constructions. However, none of the sites mentioned here would normally qualify for the name home of the gods. On the other hand, no. 9 (above), namely gudhem denoting an undefined area, a forest, a mountain, a lake etc., could perhaps have been looked upon as a home of the gods. A semantic relative could therefore be Tiveden, the border woodland between the provinces of Västergötland and Närke in central Sweden. The name seems to contain the words *ti- gods (cf. ON tívar gods ) and OSw vidher forest, hence probably with the meaning the forest where gods or supernatural beings dwell or appear. Another example could be a case found in Uppland. During the Late Iron Age, an inlet stretched northwards from Lake Mälaren (by then part of the sea) to just east of the city of Enköping. In the inner part of this inlet is a small island, today called Enhelga, which during the Middle Ages was called OSw Øn helga the holy island. Karin Calissendorff (1964, ) has made the plausible assumption that the neighbouring hamlet of Gåde (< Gudhø the island of the god ) bears the older name of the island. Gods were therefore assumed to dwell or appear on this island. Interestingly, this island is known to have had an assembly function as well. The problem is that these two examples are well-defined natural features a forest and an island whereas the Gudhem/Gudme names have no such clear reference points. I shall end by presenting a location in the pre-christian Scandinavian landscape that might well deserve the name home of the gods, namely the island of Selaön in Lake Mälaren in central Sweden (Fig. 1; see Brink 1999), despite the fact that it is outside the area in Scandinavia where we find the Gudhem names. This island is rather unique in several ways. The toponymic evidence is extraordinary and very illuminating for our purpose. In the very south, near Kolsundet, the narrow sound between Selaön and the mainland of Södermanland, and beside a small bay, is the focal site of Selaön, the prehistoric centre of power and administration. Here lie the hamlet of Tuna and the nearby royal hamlet of Husby. 3 a mountain; and that mountain Thorolf considered so holy that he decreed that no unwashed man should look at it That mountain he called the Holy Mountain and he believed that he should go there when he died, and all his kindred from the peninsula (my translation). 20

7 Mälaren + Algö Runsö Lytislunda Vittinge Kilfröslunda Överselö Klippinge Viggeby St. Fröslunda Nällsta Fröberga Lundby Åsa Ullunda Odensicke Mälsåker Karlby Opptuna Ytterselö Tuna Husby km Kolsundet Fig. 1. The island of Selaön in Lake Mälaren, central Sweden. Adjacent to the latter is Karby (< Karlabyr). Linked with this central-place complex is a former thing site, test-ified on a rune stone standing here at Kolsundet (Sö 196) 4. This physical evidence can be correlated with information in medieval documents, which indicate that the thing site for the Selebo hundare was Kolhøgha on Selaön 5. Here, therefore, in close proximity, are Tuna, Husby, Karby and a prehistoric thing site, Kolhøgha. North of Tuna lies the hamlet of Ullunda the cultic grove dedicated to Ull. A few kilometres further north is Nällsta (probably < Niærdha(r)staver) the staff or idol connected to the goddess *Niærdh. On the west side of the island is Fröslunda the cultic grove dedicated to the god Frö. Nearby is another hamlet called Fröberga the hill dedicated to the goddess Fröja. This combination of two cultic place names, close together, with a female and a male god as the first elements, is probably no coincidence. They have been interpreted as having some kind of function in a pagan fertility cult. In the northern part of Selaön is another Fröslunda, the cultic grove dedicated to the god Frö, with and this is extremely interesting a neighbouring farm called Lytislunda (today Janslunda), which has been interpreted as the cultic grove connected to lytir (a pagan cult leader) (Elmevik 1990; 2003). 4 hikifriþr lit r-isa stin þna iftir ayulf faþur sin auk staf ayulf- kiarþi þat ausþiki hiuk asur ifnti kina uistr Ingifriðr let ræisa stæin þenna æftir Øyulf, faður sinn, ok staf. Øyulfr gærði þæt austþingi (?). Hiogg Assurr. Æfndi (?) Ginna vestr Ingefrid raised this stone after Öulv, her father, and staff. Öulv made this thing in the east (?). Asurr carved (the runes). Ginna made (thing site) in the west (?) ( 5 E.g. a rættom thingxdagh oc thingxstadh a Sylbo thinge a Kulløghum (1419; SR 3 p. 170). 21

8 Finally, in the eastern part of the island, is the famous castle and noble manor of Mälsåker. It is tempting to see Odensicke (< Odens-eke) the cultic oak grove dedicated to the god Oden and the lost Odensåker arable land connected with the god Oden in combination with the noble estate of Mälsåker (< Mælis-akr), which probably had a long tradition as an aristocratic residence, and to assume that Óðinn was the pagan god of the kings and chieftains of the community. This, of course, is merely a fantasy hypothesis, impossible to prove. This situation, with all these theophoric place names, including in principle all the major gods and goddesses worshipped in this part of Scandinavia as well as an obviously pagan cult leader, all concentrated on an single island, is, to my knowledge, totally unique: the island may well earn a designation as a sacral landscape, a home of the gods. But it is an island, not a single settlement or a farm, and it is in no way connected with the name Gudhem. Among the known Gudhem places, is at least one that may have some similarity to Selaön, namely Gudme on Fyn, which is obviously a central place, with a hall and the three hills Gudbjerg, Albjerg and Gjaldbjerg (see Kousgård Sørensen 1985; cf. Andersson 2007, ). This unique combination of names and archaeological finds has, of course, tempted many scholars to formulate more or less imaginative interpretations. Here, the name Gudme seems no doubt to have a cultic background, but it must be finds are contemporary. Gudum at Botne, Vestfold, also has an interesting setting with some notable place names: Guåker, Gurann, Gullhaug and Hvitstein, but these place names are spread throughout the parish, with no obvious contact between them. Magnus Olsen s interpretation of a Kultgebiet is therefore very questionable in this case. The same applies to Gudum at Skiptveit, Østfold. The name is found in a small settlement district (grend) in the parish, which includes three farms called Lund, Lunder and Haug: these could be used as an argument for a theophoric interpretation, but this is probably not valid. Gudhem in Västergötland, has another unique quality. It is the only Gudhem name found in an unambiguous central-place complex. There is a certain similarity to Gudhem in Östergötland, which has some interesting central-place and theophoric place names in the vicinity but, again, it is very difficult to understand their connections and background. Finally, there is Gudhjem on Bornholm, which has no toponymic context indicating that there is something special to be found here. However, as Anne Nørgaard Jørgensen shows (in this vol.), the archaeological context is most interesting (cf. Bent Jørgensen in the present book for a more thorough analysis of the Danish names of that kind). To sum up, the name Gudhem is very elusive and I cannot find any potentially obvious interpretation. It is not even possible to decide whether the names denoted a specific settlement, a farm, an area, or a district. Not even the landscape analysis is of any assistance: there are no common features or recurring elements at these sites. Many of the Gudhem settlements are found at central locations within their districts but they are not focal points in the landscape, hence not the centre. For me, the Gudhem problem, both the place names and the settlements, is even more of an enigma after this analysis than before. References Amituana i 1986: T. Amituana i, Kava in Samoa. In G. and B. Deverell (eds.), Pacific Rituals. Living or Dying (Suva 1986) Andersson 2007: Th. Andersson, Zentralorte onomastisch. In: RGA 2 34 (Berlin, New York 2007) Beck 1995: H. Beck, Namenkundlich-religionsgeschichtliche Bemerkung zur Gudme-Diskussion. In: E. Marold/Ch. Zimmermann (Hrsg.), Norwestgermanisch. RGA 2 Ergbd. 13 (Berlin, New York 1995) Brink 1991: S. Brink, Iakttagelser rörande namnen på -hem i Sverige. In: Heidersskrift til Nils Hallan på 65-årsdagen 13. desember 1991 (Oslo 1991)

9 Brink 1995: S. Brink, Home. The term and the concept as seen from a linguistic and settlement-historical point of view. In: C. N. Benjamin/D. Stea (eds.), The Concept of Home: An Interdisciplinary View (London, New York, Sydney 1995) Brink 1997: S. Brink, Onomastiska utbredningskartor. Vad står de för egentligen? In: L.-E. Edlund (ed.), Kartan i kulturforskningens tjänst. Kulturens frontlinjer 2 (Umeå 1997) Brink 1999: S. Brink, Social order in the early Scandinavian landscape. In: C. Fabech/J. Ringtved (eds.), Settlement and Landscape (Århus 1999) Brink 2001: S. Brink, Mythologizing Landscape. Place and Space of Cult and Myth. In: M. Stausberg (Hrsg.), Kontinuitäten und Brüche in der Religionsgeschichte. RGA 2 Ergbd. 31 (Berlin, New York 2001) Brink 2007: S. Brink, How Uniform was the Old Norse Religion? In: J. Quinn/K. Heslop/T. Wills (eds.), Learning and Understanding in the Old Norse World. Medieval Texts and Cultures of Northern Europe 18 (Turnhout 2007) Calissendorff 1964: K. Calissendorff, Helgö. Namn och bygd 52, 1964, Egils saga, ed. by Bjarni Einarsson (London 2003). Elmevik 1990: L. Elmevik, Aschw. Lytis- in Ortsnamen. Ein kultisches Element oder ein profanes? In: T. Ahlbäck (ed.), Old Norse and Finnish Religions and Cultic Place-Names. Scripta Instituti Donneriani Aboensis 13 (Stockholm 1990) Elmevik 2003: L. Elmevik, En svensk ortnamnsgrupp och en hednisk prästtitel. Ortnamnssällskapets i Uppsala årsskrift (2003), Kousgård Sørensen 1985: J. Kousgård Sørensen, Gudhem. Frühmittelalterl. Stud. 19, 1985, Jørgensen 2009: L. Jørgensen, Pre-Christian cult at aristocratic residences and settlement complexes in southern Scandinavia in the 3 rd 10 th centuries AD. In: U. von Freeden/H.Friesinger/E. Wamers (Hrsg.), Glaube, Kult und Herrschaft. Phänomene des Religiösen im 1. Jahrtausend n. Chr. Akten des 59. Internationalen Sachsensymposiums. Koll. Vor- und Frühgesch. (Bonn 2009) Sahlgren 1954: J. Sahlgren, Gudhem. Svenska landsmål och svenskt folkliv 76/77, , 1 7. Sahlgren 1958: J. Sahlgren, Review of: Danmarks stednavne 13. Namn och bygd 46, 1958, Samnordisk runtextdatabas = SD = Svenskt diplomatarium: Diplomatarium Suecana 1 (Stockholm 1829ff.). SR = Sveriges runinskrifter 1, publ. by Kungl. Vitterhets-, historie- och antikvitetsakademien (Stockholm 1900ff.). Tuan 1974: Y.-F. Tuan, Topophilia. A Study Of Environmental Perception, Attitudes, And Values (Englewood Cliffs, New Jersey 1974). Tuan 1977: Y.-F. Tuan, Space and place. The perspective of experience (London 1977). Vikstrand 2001: P. Vikstrand, Gudarnas platser. Förkristna sakrala ortnamn i Mälarlandskapen. Acta Academiae Regiae Gustavi Adolphi 77 (Uppsala 2001). Wilson et al. 1990a: J: Wilson/T. Aue Davis/T. O Regan, Ngâ Tohu Pûmahara. The Survey Pegs of the Past. Understanding Mâori Place Names (Wellington 1990). Wilson et al. 1990b: J. Wilson/T. Aue Davis/T. O Regan, He Korero Pûrâkau Mo Ngâ Taunahanahatanga A Ngâ Tûpuna. Place Names of the Ancestors. A Maori Oral History Atlas (Wellington 1990). Prof. Dr. Stefan Brink Centre for Scandinavian Studies University of Aberdeen 24 High Street GB AB24 3EB Aberdeen s.brink@abdn.ac.uk 23